Report

Natural Selection and the Reinforcement of Mate Recognition

See allHide authors and affiliations

Science  20 Oct 2000:
Vol. 290, Issue 5491, pp. 519-521
DOI: 10.1126/science.290.5491.519

Abstract

Natural selection on mate recognition may often contribute to speciation, resulting in reproductive character displacement. Field populations of Drosophila serrata display reproductive character displacement in cuticular hydrocarbons when sympatric withDrosophila birchii. We exposed field sympatric and allopatric populations of D. serrata to experimental sympatry with D. birchii for nine generations. Cuticular hydrocarbons of field allopatric D. serrata populations evolved to resemble the field sympatric populations, whereas field sympatric D. serrata populations remained unchanged. Our experiment indicates that natural selection on mate recognition resulted in the field pattern of reproductive character displacement.

A direct role for natural selection in the generation of reproductive isolation during speciation is highly controversial (1–6). Natural selection may rapidly increase divergence in mate recognition between sympatric populations of speciating animal taxa by selecting against hybridization between heterotypic individuals. The reinforcement of mate recognition by this process will result in the pattern of reproductive character displacement (3), where sympatric populations of closely related species have diverged in mate recognition to a greater extent than allopatric populations.

Reproductive character displacement has been found in a diverse range of taxa (3–9). However, the pattern of reproductive character displacement does not constitute direct evidence for natural selection reinforcing mate recognition (10) because it may be caused by a number of other processes (3, 6). In particular, it has yet to be shown that a trait which displays reproductive character displacement in the field is heritable and responds to natural selection on mate recognition within the context of an experiment that excludes other possible causes of the displacement. Here, we use a natural selection experiment (11) to demonstrate the role of natural selection in the generation of a field pattern of reproductive character displacement.

Drosophila serrata and D. birchiihave different but overlapping distributions and habitat associations along the east coast of Australia (12). The two species are sexually isolated by pheromones composed of cuticular hydrocarbons (CHCs) (13, 14) and hybrids are viable and fertile (12, 13). Species in which a single reproductive trait is largely responsible for mate recognition provide the best systems for studying reproductive character displacement (15). In this species pair, mate recognition and CHC profile are highly genetically correlated, with genetic variation in CHCs accounting for over 70% of the genetic variation in mate recognition between the two species (14). Selection on mate recognition is therefore likely to result in changes in CHC profile. The CHCs of D. serrata display reproductive character displacement, changing abruptly at the sympatry-allopatry border (Fig. 1), although the pattern is still confounded with geographic distribution as in many other examples (16).

Figure 1

Reproductive character displacement in the cuticular hydrocarbons of D. serrata. The map of eastern Australia shows the distributions of D. serrata and D. birchii as well as the positions of three sympatric (▪) and five allopatric (•) populations of D. serrata. Population means are from the first two canonical variates from a canonical discriminant analysis conducted using the CHC data from the control populations from the selection experiment plus the two additional populations (4 and 5).

To test whether natural selection on mate recognition generated the reproductive character displacement in the CHCs of natural populations of D. serrata, we exposed three field sympatric and three field allopatric populations of D. serrata to experimental sympatry with D. birchii for nine generations (17). We predicted that if the field pattern of reproductive character displacement was the result of natural selection on mate recognition in sympatry, the CHCs of field allopatric populations would evolve in experimental sympatry, whereas those of field sympatric populations would not. The CHCs of experimental sympatry and control lines were assayed after selection ended (18). A canonical discriminant analysis (19) was used to display the relation between the experimental populations in multivariate CHC space (Fig. 2). The first two canonical variates (CV1 and CV2), accounting for 94.9% and 2.4% of the variation, respectively, were analyzed in univariate split-plot analyses of variance (20). The interaction between the treatment of experimental sympatry and whether the populations were originally allopatric or sympatric in the field tested if evolution in experimental sympatry had occured differentially in field allopatric and sympatric populations. Paired t tests were used to determine if the presence of an interaction was a consequence of evolution in field allopatric populations, rather than in field sympatric populations.

Figure 2

The effect of natural selection on the cuticular hydrocarbons of D. serrata after nine generations of experimental sympatry with D. birchii. Evolutionary responses are indicated from control populations (closed symbols) to experimental sympatry populations (open symbols) of field allopatric populations (circles connected by solid arrow) and field sympatric populations (squares connected by dashed arrow). Numbers refer to localities given in Fig. 1.

Drosophila serrata females from field allopatric populations exposed to experimental sympatry evolved toward the field sympatric control populations in multivariate CHC space (Fig. 2). Univariate analysis of variance (ANOVA) conducted on female CV1 and CV2 indicated interactions between the treatment of experimental sympatry and whether the population was allopatric or sympatric in the field [CV1, F(1,4) = 8.199, P = 0.046; CV2,F(1,4) = 7.519, P = 0.052]. Evolution occurred in field allopatric populations (paired ttests: CV1, t2 = 4.273, P = 0.051; CV2, t2 = –5.086, P= 0.037), but not in field sympatric populations (CV1,t2 = –1.609, P = 0.249; CV2, t2 = –0.525, P = 0.652). Drosophila serrata males from two of the three field allopatric populations exposed to experimental sympatry also evolved toward the field sympatric control populations. The males of the third field allopatric population (Wollongong) evolved in the same direction along CV1 as the first two populations but in the opposite direction on CV2 (Fig. 2). Univariate ANOVA conducted on male CV1 indicated an interaction between exposure to experimental sympatry and whether the population was allopatric or sympatric in the field [F(1,4) = 7.056, P = 0.057], which was a consequence of evolution in field allopatric populations (t2 = –8.628, P = 0.013) and no change in field sympatric populations (t2 = –0.362, P = 0.752). Male CV2 displayed no interaction as a consequence of the Wollongong population.

Experimental sympatry resulted in the evolution of the CHCs of both sexes of D. serrata originating from field allopatric populations. In general, the response of both sexes was to evolve toward the field sympatric control populations. The CHCs of field sympatric populations did not respond to the presence of D. birchii in experimental sympatry, indicating that the reproductive character displacement displayed in Fig. 1 was a consequence of natural selection on mate recognition under field conditions.

Reproductive character displacement evolved within nine generations, indicating that there was strong selection on mate recognition. To determine whether selection on mate recognition operated either prezygotically during courtship or postzygotically after hybrids were formed, we conducted two experiments. First, the frequency of successful hybridization in our experimental sympatry treatments was determined (21) and was found to be very low, suggesting that little selection was applied postzygotically. Second, prezygotic selection was assessed by determining the efficiency ofD. serrata males in inseminating D. serratafemales in experimental sympatry (22). Experimental sympatry differentially affected the efficiency of field sympatric and field allopatric males in inseminating females [F(1,2) = 64.39, P = 0.015] (Fig. 3). Field allopatric males inseminated significantly fewer females in experimental sympatry than in the allopatric controls (paired t test:t 2 = 5.179, P = 0.035), with nearly 50% fewer D. serrata females inseminated in the presence of D. birchii. In contrast, the number of females inseminated by field sympatric males was unaffected by experimental sympatry (t 2 = 1.220, P = 0.347). Selection on mate recognition therefore operated during courtship, rather than after the production of hybrid individuals with low fitness, to generate the reproductive character displacement in CHC profile.

Figure 3

The effect of experimental sympatry on the efficiency of field sympatric and allopatric D. serrata males in inseminating D. serrata females. Means and 95% confidence intervals are based on the three field allopatric (•) and sympatric (▪) populations.

Although natural selection in our laboratory environment operated during courtship, prezygotic selection for mate recognition may not be as intense under natural conditions if adult densities are lower and the larval substrate is not limited to a single patch (bottle). The form that selection takes under field conditions to result in reproductive character displacement in this system remains to be evaluated. Furthermore, this experiment does not indicate whether natural selection on mate recognition in sympatry was a component of the historical speciation event between D. serrata andD. birchii, but rather indicates how rapidly the mate recognition system of populations may evolve when confronted with the presence of a closely related group. The large number of cases of apparent reproductive character displacement in field populations across diverse taxonomic groups (3–9) suggests that natural selection on mate recognition may be a major component of the evolution of mate recognition in many animals (1).

  • * To whom correspondence should be addressed: E-mail: MHiggie{at}zoology.uq.edu.au

REFERENCES AND NOTES

View Abstract

Stay Connected to Science

Navigate This Article